A refined ground-borne noise prediction methodology for railway traffic in tunnels in bedrock

Doctoral thesis, 2025

The expansion of railway networks has significantly improved transportation efficiency but has also led to increased noise and vibration, particularly affecting residential areas near underground infrastructure. Ground-borne noise from rail traffic in tunnels can impact human health, structural integrity, and overall environmental quality. To support effective mitigation and infrastructure planning, accurate and reliable prediction models are needed.

This study developed a model and methodology for predicting ground-borne noise from railway traffic in tunnels, tailored for Swedish conditions with high-quality bedrock. Existing models are often proprietary, with limited data available for Swedish bedrock conditions, and the handling of uncertainties is often insufficiently explained. This work addresses these gaps by developing a structured, multi-stage model to support the Swedish Transport Administration projects. The methodology consists of three stages, location, planning, and construction, each adapted to the level of data available. The model is valid up to 1~kHz and incorporates a source term along with correction terms for train speed, distance attenuation, ground-to-building coupling, vibration transmission through structures, and room acoustics. Statistical uncertainty is included for each term, ensuring robust predictions.

The model is based on field measurements from the Gårda tunnel in Gothenburg and the Åsa tunnel in Varberg. To enhance understanding, numerical simulations were also conducted to investigate the effects of cracked bedrock zones and tunnel structures on vibration propagation. The simulations showed that the cracked zone causes frequency-dependent attenuation beyond the zone and amplification on the source side under idealized conditions. Tunnel structures were found to reduce vibration levels above the tunnel and introduce fluctuations at higher frequencies.

Additional field tests were conducted in a tunnel under construction using both shaker and hydraulic hammer excitations to further refine the model by assessing vibration transfer to nearby buildings. While these tests allowed for a comparison between excitation sources, no significant vibration was detected at the house level.

As a result, a methodology and detailed prediction model is proposed for ground-borne noise assessment in Swedish Transport Administration projects.