MICROSEISMIC MONITORING OF GEOTHERMAL ENERGY EXPLOITATION OPERATIONS USING DISTRIBUTED ACOUSTIC SENSING (DAS)

A widespread application of technologies involving fluid injection in the subsurface will be key in the next decades to reduce the carbon footprint of industrial activities. Carbon Capture and Storage (CCS) and Enhanced Geothermal Systems (EGS) play a central role in reaching this target. While CCS could reduce the environmental impact of using conventional energy resources, EGSs have the potential to make geothermal energy a valuable clean resource no longer confined to volcanic or hydrothermal regions. However, these technologies are not free of risks, requiring detailed site characterization and monitoring to ensure operational safety and public acceptance. In particular, both CCS and EGS are known to play a role in inducing earthquakes; while EGS may induce felt or damaging earthquakes that can be a danger for the population, induced seismicity in CCS sites can be an indicator of loss of integrity of geological safeguards, and potential CO2 leaks in the atmosphere. This thesis aims to improve the sustainability and safety of CO2 storage operations and mitigate the risks of geothermal energy exploitation through an innovative seismic data acquisition technology based on the use of fiber-optics sensors and known with the name of Distributed Acoustic Sensing (DAS). Fiber-optic cables can be easily deployed in deep boreholes or even in the same wells used for injection/production, reducing the costs of the monitoring infrastructure. Because of this advantage, the distance between the detecting sensor and the induced seismicity can be minimized, hence maximizing our capability to detect a large number of tiny earthquakes. While DAS represents a breakthrough in observational capabilities, it also poses challenges in seismological data processing and analysis yet to be faced. Currently, the main limitation of DAS systems is the burden placed on data storage infrastructure owing to the massive volume of data they produce. In addition, the different kinds of noise affecting DAS data require specialized analysis techniques and standard seismological techniques are not yet ready to exploit the main features of DAS (e.g., the high spatial density) and manage such big data volumes. Within this thesis, we focused on the development of a waveform-based microseismic event detection method that is able to exploit the high sensor density of DAS. This method was successfully applied to microseismic data collected during fluid injection experiments at the EGS Utah-FORGE site, where it demonstrated a remarkable capability by detecting nearly twice the number of seismic events compared to the Silixa reference catalog. Additionally, we developed a seismic noise analysis method to address the literature gap on noise in fiber optic data, studying depth-related noise anomalies and identifying potential mechanical noise components affecting the DAS interrogator unit. The high sensor density enabled phase analysis of ambient noise via cross-correlation, distinguishing coherent noise from surface sources and noise generated in proximity of the fiber. Applying our methods to synthetic datasets was crucial for implementation. Managing all involved variables allows us to identify and solve potential issues, leading to more effective method development. Finally, we are developing a new tool, based on the Pyrocko library, for generating DAS synthetic waveforms. This addresses the limited availability of DAS data, enabling the community to use realistic synthetics and accelerating the development of effective industrial monitoring techniques.