Dynamic system identification and digital twins for underground infrastructure monitoring

This project investigates vibration-based methods for monitoring the structural health of deep underground tunnels throughout their lifetime. Structural Health Monitoring (SHM) based on dynamic system identification and digital twins has demonstrated significant advantages when applied to bridges, wind turbines, and high-rise buildings. Variations in dynamic characteristics - such as eigenfrequencies, mode shapes, and modal damping ratios - can indicate structural changes or damage. Along these lines, this project develops SHM methods for deep tunnels, with the HADES Underground Research Laboratory (URL) serving as the primary case study.

The monitoring strategy combines in situ vibration measurements with advanced numerical simulations. Deep underground structures are strongly influenced by complex dynamic soil–structure interactions. Vibration-based monitoring captures the coupled response of the tunnel and surrounding soil. Conventional techniques—such as visual inspection, strain gauges, and convergence measurements—often fail to detect distributed or deep-seated defects, require manual access, and offer limited spatial and temporal resolution. These methods also lack the sensitivity to detect early changes in global structural behavior. In contrast, wave-based techniques offer a complementary and more responsive alternative, though their application to tunnel–soil systems remains underexplored.

Coupled numerical models of the tunnel and surrounding soil will be developed, combining finite element with boundary element and perfectly matched layer formulations. These models will be used to compute modal characteristics, as well as dispersion and attenuation curves of guided waves along the tunnel. The models will simulate various environmental and damage scenarios, such as thermal loading and local changes in soil properties, and be used for parametric and sensitivity studies.

Forced and ambient vibration experiments will be conducted in the HADES URL to validate the models. Hammer impact tests will identify dispersion and attenuation curves and mode shapes, while long-term ambient monitoring will capture operational vibration patterns due to anthropogenic and natural excitation. The models will help interpret how different damage scenarios affect the ambient vibration signature. Both black-box and white-box damage identification algorithms will be assessed.

The resulting digital twin will serve as the foundation for a permanent SHM system for the HADES tunnel. The developed methodology will also be applicable to future underground infrastructures such as geological disposal facilities, the Einstein Telescope and railway tunnels.