Low frequency vibrational behaviour in concrete structures; exploring structural reverberation time and distance based attenuation trends across building floors

Abstract

This thesis investigates low-frequency structure-borne sound propagation in concrete floor slabs across multiple floors of a building. The study aims to assess whether a relationship exists between reverberation time, distance-dependent vibration level decay, and group velocity. Both in-situ measurements and numerical simulations using Comsol Multiphysics were conducted to analyse structural vibrational behaviour and estimate parameters such as propagation velocity, reverberation time, and attenuation per distance (dB/m) in the 50–160 Hz third-octave band range. The structural impulse responses were post-processed through backward integration and time-delay estimation techniques, allowing for comparison between simulated and real building conditions. Five analytical propagation models combining geometrical spreading and damping were fitted to the data using Matlab functions. The results from both simulations and measurements suggest that no consistent or global link could be established for the reverberation time (T60), and therefore no clear proportional relationship was found between T60 and the vibrational level decay over distance (La). However, some tendencies toward a proportional relationship between La per distance and the effective group velocity ceff were observed. Notably, the velocity trends diverged between simulation and measurement, indicating potential differences in wave type dominance over distance for the measured data and the simulated results. The findings reveal the limitations of modeling large, complex structures due to computational constraints and emphasise the importance of local geometry and boundary conditions in structural acoustic behaviour. Although a complete predictive model could not be established, this work provides a framework for assessing energy-based decay behaviour in solid structures. It proposes directions for future development of building-scale low-frequency vibration prediction models.