Studio dei principali aspetti che influenzano l'efficienza della roccia di copertura per la valutazione della sicurezza dello stoccaggio sotterraneo di gas tramite algoritmi di Machine Learning. = Investigation of the main aspects affecting the sealing efficiency of cap rock for the safety assessment of underground gas storage via Machine Learning algorithms.

Underground gas storage plays a pivotal role in ensuring energy security, carbon management, and industrial applications. The efficiency and long-term stability of gas containment are governed by complex interactions between geological formations and fluid properties, particularly breakthrough and snap-off pressures. These pressures influence the displacement of fluids within porous media and are integral to optimizing storage capacity and preventing leakage. While lithological factors have been extensively studied in relation to these pressures, the role of mineralogical composition remains underexplored. This research seeks to bridge this gap by developing a predictive model that correlates breakthrough pressure, snap-off pressure, and key reservoir parameters, including pressure, temperature, porosity, permeability, interfacial tension, and contact angle. The study leverages a comprehensive dataset compiled from multiple sources, encompassing laboratory experiments and geological formations worldwide. Advanced experimental methodologies, such as breakthrough pressure tests and mercury intrusion porosimetry, are employed to assess fluid displacement dynamics. Furthermore, state-of-the-art machine learning techniques, specifically K-Nearest Neighbors (KNN) and CatBoost, are utilized to analyze the complex, nonlinear relationships between reservoir characteristics and pressure behavior. These models enhance the accuracy of pressure predictions in subsurface storage systems, offering valuable insights into fluid retention mechanisms. A focal point of this research is the Opalinus Clay formation, widely recognized for its effectiveness as a cap rock in geological storage applications. The study systematically examines the influence of mineralogical composition, pore structure, and wettability on capillary pressure and fluid migration. The findings highlight the significant impact of mineralogical variations on pressure dynamics, emphasizing their implications for hydrocarbon recovery, CO₂ sequestration, and hydrogen storage. By integrating geological expertise with data-driven modeling, this research advances the understanding of subsurface fluid storage and enhances predictive capabilities for pressure behavior in geological formations. The outcomes contribute to the development of more robust pressure estimation models, thereby optimizing gas storage strategies and reinforcing the safety and reliability of long-term CO₂ and H₂ sequestration initiatives.