Shift2Rail logo A body of the European Union

Cross-Cutting Activities

IP Coordinator: Richard French - BT


Total Project Value:
€ 949 999,50
from 01/11/2020 to 31/10/2022
S2R (Of H2020) co-funding:
€ 949 999,50
Pascal Bouvet
Complementary projects:
Project website:


The overall goal of SILVARSTAR is to provide the railway community with proven software tools and methodologies to assess the noise and vibration environmental impact of railway traffic on a system level.

The first overall objective of SILVARSTAR is to provide the railway community with a commonly accepted, practical and validated methodology and a user-friendly prediction tool for ground vibration impact studies. This tool will be used for environmental impact assessment of new or upgraded railways on a system level. It will provide access to ground vibration predictions to a wider range of suitably qualified engineers and will facilitate project planning and implementation by improved simulation processes.

The second overall objective of SILVARSTAR is to develop a fully functional system for auralisation and visualisation based on physically correct synthesised (exterior and interior) railway noise, providing interfaces with Virtual Reality visualisation software. This system will facilitate communication with the public, decision makers and designers through virtual experience before delivery of projects.

In order to fulfil these overall objectives, the following specific scientific and technical objectives are addressed:

1. To develop a modular hybrid methodology for ground vibration prediction.

SILVARSTAR will build on recent developments of hybrid models that combine the advantages of both numerical and experimental approaches. The general framework recommended in the ISO 14837-1:2005 standard will be followed where the vibration level in a building is written as the product of a source, a propagation and a receiver term; each of these terms is frequency-dependent and can be represented by numerical prediction or by experimental data.

The vibration source is the train-track interaction. The proposed method will allow the properties of the vehicles, the track, and the underlying ground to be fully taken into account, as well as excitation in terms of the track and wheel surface unevenness.

The propagation path is highly complex, depending on the local soil properties. The use of measured transfer functions can directly take account of on-site complexity, but cannot allow for situations that do not yet exist (e.g. the soil will change due to the construction of the railway infrastructure). In this way, numerical approaches are also useful.

Finally, at the receiver noise and vibration levels depend on the building structure and its foundation type. The large variety of structures is generally addressed with statistical transfer functions that depend on overall building type and use. The same approach will be adopted and will be coupled to the two previous terms.

2. To enable transposition of vibration emission data from one situation to another.

Ground vibration predictions often rely on source data, relating to the train-track interaction obtained at one location, and transmission path data relating to another. Such vibration data must therefore be easily transposed from one situation to another. By its modular nature, the method to be developed in SILVARSTAR will allow the user to change the source parameters (train, track, roughness excitation level), the propagation path (soil), and the receiver (building). The validity of such a transposition approach will be verified by comparison with well-documented reference cases.

3. To define and validate a track-independent vehicle indicator.

In the field of noise, standard test methods are available to define the noise emission of individual vehicles, which are largely independent of the track (ISO 3095:2013). For ground vibration, emission levels are strongly dependent on the track and ground properties so that measurements at different sites cannot be directly compared. The purpose is to develop a suitably reliable track-independent vehicle indicator that can be used to identify ground-borne noise friendly vehicles. This will make use of the transposition techniques implemented in SILVARSTAR and will be validated by comparison with advanced calculation methods.

4. To implement the prediction methodology in user-friendly software with a Graphical User Interface.

The user-friendly software with a Graphical User Interface (GUI) must be adapted for project engineers, with suitable knowledge of noise and vibration, but without a need for highly specialised technical expertise. The input data should be compatible with data typically available in railway projects (geological input, track-building distance, etc.), and the output data must be compatible with descriptors defined in the standards (ISO 2631-2:2003, BS 6472-1:2008, DIN 4150-2:1999) and with Geographical Information System (GIS) visualisation similar to noise mapping. The input structure should include as much commonality as possible with noise mapping, for example the definition of train types, track categories, traffic density, etc.

The user will be able to incorporate input data from external numerical software or measurements to cover situations that are not already included in the tool and to allow more detailed assessment of specific situations.

5. To synthesise railway noise in high quality.

The objective is to develop tools to assess railway noise scenarios by audible experiences. The developments will aim at simulating train pass-bys (exterior noise), as well as vehicle acoustic comfort (interior noise).
State-of-the-art auralisation methods will be adapted to the railway sector, in order to render realistic and high-quality sounds. Existing methods from the H2020 project DESTINATE will be extended to account for specific railway noise characteristics and to allow for the required application cases. Physics-based models developed by the SILVARSTAR partners will be further developed and integrated to ensure physically correct synthesised railway noise in high quality.
The latest information on noise sources and noise mitigation measures will be incorporated into the simulation process. The required data will come from existing measurement results (consortium and external inputs), from numerical simulations using state-of-the-art tools and possibly from new measurements to be carried out.

6. To couple auralisation to visualisation to create audio-visual Virtual Reality and ready-to-use software.

The objective is to provide tangible immersive experiences for users, i.e. to see the environmental change at the same time as to hear it. These audio-visual simulations shall be physics-based and perceived as plausible.

To that end, the newly developed auralisation tools will be coupled to existing Virtual Reality visualisation software. Adequate interfaces will be established to link the created spatial sound to different visualisation tools. Coupling to real-time 3D visualisation will allow for immersive audio-visual experiences of the user. This will be demonstrated by developing a Virtual Reality application using state-of-the-art software and hardware from the gaming industry.

The final aim is to integrate the developed methodologies and auralisation models coupled to visualisation into software that can be widely used for future demonstrations. To do so, new software will be implemented, validated and tested within SILVARSTAR and finally distributed as ready-to-use freeware tools.

Results and Publications

D1.1 State-of-the-art and concept of vibration prediction tool


* D1.2 Description of the vibration prediction tool


* D1.3 Validation of the prototype vibration prediction tool against documented cases


D2.1 Database for vibration emission, ground transmission and building transfer functions


D4.1 Methodology for auralisation and visualisation of railway noise


SILVARSTAR Newsletter 2022


* Please note that this/these deliverable(s) is/are undergoing S2R JU review and acceptance processes.

All deliverables, results and publications herewith provided reflects only the author's view and the S2R JU is not responsible for any use that may be made of the information it contains.

This project has received funding from the European Union’s Horizon 2020 research and innovation programme under grant agreement No: 101015442