Seismic response analysis of unstable rock masses through passive seismic monitoring and numerical modeling at laboratory and field scales

This thesis research project was conducted as part of the 24-month PRIN initiative titled "THERMO- and HYDRO-mechanical monitoring and modeling of jointed ROCKS: a national SITE LABORATORY network (THEROCKLAB)." Starting in January 2024, THEROCKLAB involves multiple universities, including Politecnico di Torino. Its objective is to employ multi-sensor integrated monitoring strategies and numerical modeling to study thermo- and hydro-mechanical processes in fractured rock masses for early warning applications. The project covers five unstable rock masses across the Alps and the Apennines: the field-scale data of this thesis is focused on the monitoring from AcutoFieldLab in Acuto, Lazio. The research project focuses on the integration of monitoring data with numerical models to enhance the understanding of rock mass behavior and on the possibility of using resonance frequency and seismic velocity changes as early warning indicators. In this way, the research aims to contribute to the development of more effective monitoring and early warning systems for rock mass failures, with significant implications for managing risks associated with unstable rock masses in regions prone to geological hazards, particularly in the context of climate change. The research was conducted at both laboratory and field scales, combining field monitoring, laboratory testing, and FEM numerical simulations. In the laboratory, UPV testing and photogrammetry were employed to characterize the elastic properties of two rock samples with different shapes: Sample A (column-shaped) and Sample B (slab-shaped). Acoustic emission triaxial tests were performed on these samples to identify potential eigenfrequencies. The triaxial tests demonstrated the challenge of detecting and tracking potential eigenfrequencies, especially for Sample B. Additionally, the tests revealed the presence of systematic instrumental noise at 2000 Hz and 4000 Hz, indicating the necessity of designing future samples with resonance frequencies outside these ranges. Numerical simulations facilitated the interpretation of experimental results and the design of future rock samples. Eigenfrequency simulation of the rock samples provided reference values for comparison with triaxial test results. Moreover, H/V spectral ratio simulations confirmed that the greater resonant response of the column-shaped sample compared to the slab-shaped sample. Furthermore, the simulations evaluated the influence of cut length on eigenfrequencies and eigenmodes in jointed samples. This led to the recommendation of a column-shaped rock sample with a subvertical cut for future tests, as this avoids the critical frequency ranges affected by instrumental noise. The field data collected between July 1 and July 31, 2022, from AcutoFieldLab underwent spectral analysis and cross-correlation in order to identify potential natural resonance frequencies and relative velocity changes. The cross-correlation analysis yielded unsatisfactory results, indicating that the proposed methodology of tracking dV/V for early warning applications is not suitable in this case. A noteworthy discovery was the identification of a potential eigenfrequency of the unstable block at approximately 75 Hz. This frequency exhibited clear oscillatory behavior in phase with temperature changes, suggesting its role as an instability indicator. Nevertheless, further investigation over extended periods and across different seasons is necessary to validate this hypothesis.